v0.1.0 - Initial MVP

Agent Context Doctor v0.1.0

Initial public release of Agent Context Doctor: a doctor-first CLI for keeping AI coding agent context files clean, minimal, consistent, and safe.

Highlights

• Scans AGENTS.md, CLAUDE.md, .cursor/rules/*.mdc, .github/copilot-instructions.md, README/docs, .env.example, .gitignore, package.json, and GitHub workflows.
• Provides deterministic checks for context drift, command mismatch, context bloat, secret boundary issues, maintainer workflow evidence, and Japanese-language guidance.
• Adds text, Japanese text, and JSON report output.
• Supports scan, score, fix --dry-run, fix --apply, sync --dry-run, and ci commands.
• Includes a first safe autofix: append .env to an existing .gitignore when needed, without reading or printing .env contents.
• Ships npm package agent-context-doctor with agent-context-doctor and acd bin aliases.
• Includes a composite GitHub Action entrypoint and CI usage docs.

Install

npx agent-context-doctor@latest scan .
npx agent-context-doctor@latest scan . --locale ja
npx agent-context-doctor@latest ci . --min-score 80

Verification

• npm package: agent-context-doctor@0.1.0
• Git tag: v0.1.0
• GitHub Actions CI: passing
• Self-scan score: 100/100

Notes

This is an early MVP. The project intentionally starts with deterministic checks and safe dry-run-first behavior rather than broad LLM-assisted rewriting.